Predominance of Clostridium difficile Ribotypes 017 and 078 among Toxigenic Clinical Isolates in Southern Taiwan

Ribotypes and toxin genotypes of clinical C. difficile isolates in Taiwan are rarely reported. A prospective surveillance study from January 2011 to January 2013 was conducted at the medical wards of a district hospital in southern Taiwan. Of the first toxigenic isolates from 120 patients, 68 (56.7%) of 120 isolates possessed both tcdA and tcdB. Of 52 (43.3%) with tcdB and truncated tcdA (tcdA-/tcdB+), all were ribotype 017 and none had binary toxin or tcdC deletion. Eighteen (15%) toxigenic isolates harbored binary toxins (cdtA and cdtB) and all had tcdC deletion, including Δ39 (C184T) deletion (14 isolates), Δ18 in-frame deletion (3 isolates), and Δ18 (Δ117A) deletion (1 isolate). Eleven of 14 isolates with Δ39 (C184T) deletion belonged to the ribotype 078 family, including ribotype 127 (6 isolates), ribotype 126 (4 isolates), and ribotype 078 (1 isolate). Among 8 patients with consecutive C. difficile isolates, these isolates from 6 (75%) patients were identical, irrespective of the presence or absence of diarrhea, suggestive of persistent fecal carriage or colonization. In conclusion in southern Taiwan, ribotype 017 isolates with a tcdA-/tcdB+ genotype were not uncommon and of C. difficile isolates with binary toxin, the ribotype 078 family was predominant.


Introduction
Clostridium difficile is the leading cause of nosocomial diarrhea with an increase in the incidence of sporadic outbreaks causing severe and fatal infections since the beginning of the century [1]. Most alarming is the outbreak of C. difficile infections (CDIs) in Quebec, Canada in 2003. During the outbreak involving 1,703 patients, CDI was the attributable cause of death in 117 (6.9%) cases and a contributing factor in additional 127 (7.5%) deaths [2]. The hypervirulent strain had been assigned as the North American pulse-field type 1 (NAP1), restriction endonuclease analysis (REA) group BI, and polymerase chain reaction (PCR) ribotype 027 (sometimes referred to as BI/NAP1/027). Three bacterial factors have been found in the epidemic C. difficile strain, including in vitro increased production of toxin A and B, fluoroquinolone resistance, and production of binary toxin [1]. Toxin A and B are transcribed from a pathogenicity locus comprised of five genes: tcdA (toxin A), tcdB (toxin B), and three regulatory genes. One of the latter, tcdC, is a negative regulator of toxin production. Binary toxin is transcribed from cdtA and cdtB [1].
Genetic relationship of the C. difficile isolates causing colonization, infection, or recurrence in the same individual remains variable in several studies. Among 20 recurrent cases, Oka found 16 (80%) cases were identical between the strains at initial infection and at recurrence [11]. Nevertheless Barbut et al. reported that of the strains from 93 hospitalized patients with recurrent CDI between 1994 and 1997, 48.4% of clinical recurrences were caused by different strains compared with initial strains [12]. However, the question of whether the initial colonized C. difficile strain was the same as or different from the strain causing subsequent infection was not answered.
In our previously published data, we reported the clinical impact and risk factor of C. difficile colonization and infection in a prospective study from January 2011 to June 2012 [8, 13,14]. We analyzed the clinical C. difficile isolates during the study period and found the first hypervirulent C. difficile ribotype 126 strain in Taiwan [15]. We further extended the prospective study to January 2013. In this study we aimed to investigate the toxin gene content and ribotype distribution of C. difficile isolates with tcdC deletion collected from previous studies [13][14][15].

Study design
We collected clinical C. difficile isolates from stool culture of the prospective study from January 2011 to June 2012 as described before [8, 13,14]. Briefly a prospective investigation was conducted in the medical wards of the Tainan Hospital, Ministry of Health and Welfare, a district hospital in southern Taiwan. The prospective clinical study was further extended to January 2013. The study was approved by the institutional review board of the Tainan Hospital, Ministry of Health and Welfare, and written informed consents were obtained from enrolled patients. Patients with age of at least 20 years old, and admitting to medical wards with expected hospital stays of at least 5 days were included. Exclusion criteria were patients with previous metronidazole or oral vancomycin therapy within three months, colectomy, or CDI at admission [16][17][18]. We retrieved demographic information, laboratory data, medication history, and underlying disease from medical records. Stool samples from the patients included in the study from January 2011 to January 2013 were sent for C. difficile culture. Stool samples were plated on cycloserine-cefoxitin-fructose agar (CCFA) under anaerobic conditions. C. difficile colonization (CdC) is defined as an asymptomatic patient with the presence of C. difficile in stool and CDI as a patient with diarrhea and the detection of toxigenic C. difficile in stool. Recurrence was defined as the resurgence of clinical symptoms after cessation of antimicrobial therapy, at least 10 days after the first episode [12].

Bacterial strains
C. difficile CCUG4938T (ribotype 001 with wild type tcdC, toxinotype 0, purchased from the Culture Collections of the University of Goteborg, Sweden), a ribotype 078 strain (with a 39-bp deletion in tcdC, provided by Prof. EJ Kuijper at Leiden University Medical Center, the Netherlands), and ATCC BAA1805 (ribotype 027 with an 18-bp deletion of tcdC, purchased from American Type Culture Collection, USA), ribotype 106 and ribotype 001/072 strain (provided by Prof. Ellie JC Glodstein at UCLA, USA), were used as reference strains.

Genomic DNA
C. difficile strains were grown anaerobically in Brain Heart Infusion broth (Becton, Dickinson and Company) with 5 mg/ml yeast extract (MO BIO Laboratories, Inc.) and 0.1% L-cysteine (AMRESCO 1 ) at 37˚C for two days. After harvesting the bacteria, C. difficile genomic DNA was extracted with a genomic DNA mini kit (Geneaid, Ltd, Taiwan).

Detection of toxin genes
The extracted DNA was amplified for the 16s rDNA, tcdA, tcdB, cdtA, cdtB, and tcdC genes of C. difficile in a single multiplex PCR, as described in [19]. The strains containing a truncated tcdC profile were further examined through tcdC sequencing, as previously described in [20]. Sequencing was performed by Mission Biotech Co., Ltd. Amplification was performed with a BigDye terminator 3.1 kit (Applied Biosystems) according to the manufacturer's instructions. Capillary sequence analysis was also performed by Mission Biotech, Taiwan with an ABI 3730xl DNA sequencer (Applied Biosystems).
Sequence analysis of gyrA and gyrB was performed as previously described [21]. Briefly, the DNA region was amplified using the primer pairs, gyrA1-gyrA2 for gyrA and gyrB1-gyrB2 for gyrB. PCR products were purified and sequenced. Pairwise alignments of DNA sequences were carried out using the BLAST server of the National Center for Biotechnology Information.

PCR ribotyping
The isolates containing tcdB and determined as toxigenic strains were further examined by polymerase chain reaction (PCR) ribotyping. The ribotyping method and PCR primers used were as described previously [22]. After PCR amplification, the samples were concentrated using the Gel/PCR DNA Fragments Extraction Kit (Geneaid, Ltd, Taiwan) and separated by the QIAxcel capillary electrophoresis system (Qiagen, Hilden, Germany) using the "OM500" method and QX Alignment Marker 15 bp/3 kb (Qiagen, Hilden, Germany).

Repetitive sequence-based polymerase chain reaction (Rep-PCR)
The identity of the consecutive isolates collected beyond 28 days from the same patient was determined by Rep-PCR. Rep-PCR was performed as described by Versalovic et al. [23], and the PCR products were analyzed using the QIAxcel system.

Statistical analysis
Statistical analyses were performed by statistical software (SPSS, version 13.0). Continuous data were expressed as the means ± standard deviations. The χ 2 test or Fisher's test was used to compare categorical variables, and Student's t-test was used to compare continuous variables. A two-tailed P value of less than 0.05 was considered to be statistically significant.
Genetic relationship of consecutive toxigenic C. difficile isolates from the same individual Genetic relationship of consecutive toxigenic isolates from the same individual collected beyond 28 days was examined by Rep-PCR (Table 2). Of 8 patients, consecutive isolates were identical in 6 (75.0%) patients, irrespective of the presence or absence of diarrhea. Genetic similarity among consecutive fecal C. difficile isolates obtained from 2 patients was illustrated (Fig 2).
We assessed the seasonal distribution of toxigenic C. difficile isolates stratified by binary toxin and ribotype 078 family (S2 Fig). More C. difficile isolates with binary toxin, particularly those belonging to the ribotype 078 family, were harvested at the first and second quarters of 2011 and 2012.

Discussion
In the present study, the ribotype 078 family was dominant among toxigenic C. difficile isolates with binary toxin and tcdC deletion in Taiwan. Moreover the ribotype 078 family in Taiwan often displayed a high level of moxifloxacin resistance. Several retrospective studies in Taiwan investigated the prevalence of hypervirulent C. difficile strains, such as ribotype 027 or 078, but no hypervirulent isolates had been discovered before 2012 [7][8][9]. To our knowledge, this is the first study describing ribotype and toxin genotype distribution of clinical C. difficile isolates in Taiwan. Our study revealed that among the isolates with binary toxin, more than half belonged to the ribotype 078 family, including ribotypes 078, 126, and 127. The result was unique in that ribotype 078 has been reported from Korea and China [24][25][26], but it not the dominant ribotype. Clinical impact and origin of the dominant ribotype 078 family among binary toxin producers in Taiwan warrant further evaluations. In Hangzhou, China in 2013, the predominant C. difficile ribotypes in hospitalized cancer patients included 001, 017/1, and 017 [27]. At another hospital in Hunan, China between April 2009 and February 2010, the dominant ribotype was 017, followed by 046 and 012 [28]. Ribotypes 018, 017 and 001 were prevalent in Seoul, Korea from September 2008 to January 2010 [29]. In general, ribotype 017 isolates were commonly present in China and Korea, and here we found that this ribotype accounted for 43% of toxigenic isolates, suggestive of its  widespread in Asia countries. An international surveillance of toxigenic C. difficile isolates is essential to disclose ribotype distribution and clinical significance of toxigenic C. difficile isolates in Asia.
Of the patients in our study, identical C. difficile isolates can be obtained from the same individuals with tCdC, CDI, or recurrent CDI. The result was similar to the study conducted by Oka et al., who noted 80% were identical between the strains at initial infection and at recurrence [11]. Moreover Kim et al. found relapse rates of ribotype 017 and 018 isolates were higher than those of other ribotypes (63.6% and 63.6% vs. 22.2%, respectively) [30]. The clinical impact of C. difficile ribotypes on CDI recurrence warrants more studies. However, in our study the same C. difficile isolate persisted from CdC to CDI or from CDI to CdC, suggestive of long-term carriage of identical C. difficile strains, irrespective of the initial presence or absence of diarrhea. Furthermore, the role of asymptomatic carriers in C. difficile transmission has been supported by a recent study in which active surveillance detection and isolation of C. difficile carriers was associated with a significant decrease in the incidence of healthcare-associated CDI [31]. Clinical significance and management of asymptomatic C. difficile carriers warrant more attention.
Patients with CDI were characterized as less likely to be nursing home residents, more likely to use nasogastric tube and to have diabetes mellitus, than those with CdC. Nasogastric tube use [32,33] and diabetes mellitus [34] had been regarded as risk factors for CDI before. It is notable that CDI in our study was inversely correlated with nursing home residing, though the prevalence of C. difficile colonization or infection among residents in long-term care facilities had been increasing in recent years due to advanced age, the recipients of multiple courses or longer duration of antibiotics, and previous CDI history [35,36].
There are some limitations in our study. Firstly, we only analyzed the ribotypes of toxigenic C. difficile isolates with binary toxin and tcdC deletion and tcdA-/tcdB+ isolates. Secondly, the C. difficile isolates studied were only obtained from a district hospital in southern Taiwan. Their representativeness is of course questionable. Thirdly, it is likely that the boundary of C. difficile colonization and infection is blurred. With longer followed-up periods, the isolates of tCdC and CDI will be exchangeable. For example, the isolates initially causing colonization can later be related to CDI, and vice versa. Finally, though a total of 13 patients were excluded from the study due to expected hospital stay for less than 5 days, recent metronidazole or vancomycin therapy, colectomy, or CDI at admission, the estimated number of toxigenic C. difficile isolates obtained from them was 2 or 3. Therefore, the exclusion of these patients will not change significantly our study results.
In conclusion, ribotype 017 isolates were not uncommon in toxigenic C. difficile isolates southern Taiwan and the ribotype 078 family predominated among clinical C. difficile isolates with binary toxin.